Two-way speech recognition and dialect system

• US 6,424,935 B1
• Filed: 07/31/2000
• Issued: 07/23/2002
• Est. Priority Date: 07/31/2000
• Status: Expired due to Term

• Current Assignee: Ovonyx Memory Technology, LLC (Micron Technology, Inc.)
• Original Assignee: Micron Technology, Inc.
• Inventors: Taylor, George W.
• Primary Examiner(s): Dorvil, Richemond
• Assistant Examiner(s): Nolan, Daniel A.
• Application Number:
  US09/628,754
• Time in Patent Office: 722 Days
• Field of Search: 704/231-236, 704/10, 704/251-257
• US Class Current: 704/10
• CPC Class Codes: G10L 15/005 Language recognition
• CPC Class Codes: G10L 15/065 Adaptation
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: G10L 2015/228 of application context

Abstract

A speech-to-text conversion system. The two-way speech recognition and dialect system comprises a computer system, an attached microphone assembly, and speech-to-text conversion software. The two-way speech recognition and dialect system includes a database of dialectal characteristics and queries a user to determine their likely dialect. The system uses this determination to reduce the time for the system to reliably transcribe a user'"'"'s speech into text and to anticipate dialectal word usage. In another embodiment of the invention, the two-way speech recognition and dialect system is capable of transcribing the speech of multiple speakers while distinguishing between the different speakers and identifying the text belonging to each speaker.

Claims

1. A system for transcribing spoken words to text, the system comprising:

an audio transducer that receives the spoken words and generates speech signals corresponding thereto;

a user interface through which a user can receive and send signals;

at least one data structure containing word data correlating text representations of words to speech signals wherein the at least one data structure includes dialect parameter data that can be used to recognize selected dialects corresponding to selected users;

a processor that receives the speech signals from the audio transducer wherein the processor compares the speech signals to the word data in the at least one data structure to produce text representations corresponding to the spoken words and wherein the processor initially sends signals to the user via the user interface asking the user questions indicative of the user'"'"'s dialect to thereby induce the user to provide answer signals to the processor that are indicative of dialect parameters and wherein the processor compares the dialect parameters from the user to the dialect parameter data in the at least one data structure to determine a dialect corresponding to the user prior to comparing the speech signals to the word data in the at least one data structure to produce text representations corresponding to the spoken words.

2. The system of claim 1, wherein the at least one data structure includes phonemic characteristic data corresponding to one of a plurality of dialects and wherein the processor, upon determining a dialect corresponding to the user, uses the corresponding phonemic characteristic data to facilitate recognition of text words in the word data corresponding to the speech signals.

3. The system of claim 1, wherein the at least one data structure includes at least one lexicon of words corresponding to a selected dialect and the processor, upon determining that the user has the selected dialect, uses the at least one lexicon of words corresponding to the selected dialect to translate the speech signals into the text representations corresponding to the spoken words.

4. The system of claim 1, wherein the dialect parameters include parameters indicative of the physical characteristics of the speaker.

5. The system of claim 4, wherein the dialect parameters include the age and sex of the speaker.

6. The system of claim 1, wherein the dialect parameters include parameters indicative of the geographical origin of the speaker.

7. The system of claim 6, wherein the dialect parameters include where the speaker grew up, the native language of the speaker, where the speaker currently resides and how long the speaker has resided in the speaker'"'"'s current location.

8. The system of claim 1, wherein the dialect parameters include parameters indicative of the educational level of the speaker.

9. The system of claim 1, wherein the audio transducer comprises a microphone assembly.

10. The system of claim 1, wherein the user interface includes a monitor, keyboard, and a mouse.

11. The system of claim 1, wherein the data structure further includes a plurality of text documents corresponding to each dialect and wherein the system presents the text document corresponding to the dialect determined for the user so that the user can read the text document aloud and wherein the system compares the text generated from the reading of the text document to the actual content of the text document and performs a self-correction process to improve the accuracy of the speech-to-text transcription and wherein each text document is adapted to minimize the time required for the system to perform the self-correction process.

12. The system of claim 1, wherein the spoken words originate from multiple speakers and wherein the processor sends signals to multiple users via at least one user interface to induce the users to provide signals to the processor that are indicative of dialect parameters and wherein the processor compares the dialect parameters from the users to the dialect parameter data in the at least one data structure to determine dialects corresponding to the users prior to comparing the speech signals to the word data in the at least one data structure to produce text representations corresponding to the spoken words and wherein the processor uses the dialect parameter data in the at least one data structure to associate the received speech signals with the corresponding text representation corresponding to each speaker and wherein the processor identifies the representative text corresponding to each speaker in a distinguishable manner.

13. The system of claim 12, wherein identifying the representative text corresponding to each speaker in a distinguishable manner comprises displaying the representative text via a user interface in different font types.

14. A machine loadable software program adapted to control the operation of a processor such that the processor translates audio word signals indicative of at least one word spoken by a speaker into corresponding text signals in response to receiving the audio word signals from an audio transducer wherein the software program further induces the processor to obtain dialect parameters from the speaker prior to receiving audio word signals from the speaker by inducing the processor to ask questions of the speaker indicative of the speaker'"'"'s dialect such that the speaker provides the dialect parameters via a user input device associated with the processor and wherein the software program induces the processor to determine the speaker'"'"'s dialect from the dialect parameters prior to translating the audio word signals into corresponding text signals so as to improve the efficiency of the processor in correlating text words to received audio word signals.

15. The program of claim 14, wherein the program further receives from at least one data structure phonemic characteristic data corresponding to one of a plurality of dialects and wherein the program induces the processor, upon determining a dialect corresponding to the user, to use the corresponding phonemic characteristic data to facilitate recognition of text words in the word data corresponding to the audio word signals.

16. The program of claim 14, wherein the program further receives from at least one data structure at least one lexicon of words corresponding to a selected dialect and the program further induces the processor, upon determining that the speaker has the selected dialect, to use the at least one lexicon of words corresponding to the selected dialect to translate the speech signals into the text representations corresponding to the received audio word signals.

17. The program of claim 14, wherein the dialect parameters include parameters indicative of the physical characteristics of the speaker.

18. The program of claim 17, wherein the dialect parameters include the age and sex of the speaker.

19. The program of claim 14, wherein the dialect parameters include parameters indicative of the geographical origin of the speaker.

20. The program of claim 19, wherein the dialect parameters include where the speaker grew up, the native language of the speaker, where the speaker currently resides and how long the speaker has resided in the speaker'"'"'s current location.

21. The program of claim 14, wherein the dialect parameters include parameters indicative of the educational level of the speaker.

22. The program of claim 14, wherein the program further induces the processor to present a text document after determining the speaker'"'"'s dialect wherein the processor instructs the speaker to read the text document aloud and wherein the processor compares the text generated from the reading of the text document to the content of the text document and further makes corrections to the program'"'"'s transcription so as to improve subsequent of unknown spoken words and wherein the text document presented corresponds to the dialect determined for the speaker such that the time required for the correct is reduced.

23. A method of translating spoken words into corresponding text words, the method comprising:

obtaining dialect parameter data from a speaker by asking the speaker questions indicative of dialect and then evaluating the answers;

determining a dialect of the speaker from the dialect parameter data;

receiving spoken words from the speaker after having determined the dialect of the speaker; and

using the dialect of the speaker to facilitate correlation of the spoken word to corresponding text words.

24. The method of claim 23, wherein obtaining dialect parameter data from the speaker comprises providing signals to the speaker thereby inducing the speaker to provide response signals indicative of the dialect parameters.

25. The method of claim 23, wherein determining the dialect of the speaker comprises comparing the obtained speaker'"'"'s dialect parameter data to dialect parameter data.

26. The method of claim 23, wherein receiving spoken words from the speaker comprises transducing the speaker'"'"'s speech with a microphone.

27. The method of claim 23, wherein using the dialect of the speaker to facilitate correlation of spoken words into corresponding text comprises selecting phonemic data corresponding to the dialect to interpret audio signals as corresponding to text signals contained within a database.

28. The method of claim 27, wherein using the dialect of the speaker to facilitate correlation of spoken words into corresponding text further comprises presenting a text document corresponding to the dialect of the speaker, directing the speaker to read the text document aloud, comparing the text transcribed from the spoken text to the content of the text document, and making corrections to the correlation so that the text signals in the database are more closely matched with the audio signals.

29. A method of translating spoken words into corresponding text words, the method comprising:

obtaining dialect parameter data from a plurality of speakers by asking the plurality of speakers questions indicative of dialect and evaluating the answers;

determining the dialects of the speakers from the dialect parameter data;

receiving spoken words from the speakers after having determined the dialects of the speakers;

using the dialects of the speakers to correlate the spoken words to corresponding text words; and

using the dialects of the speakers to identify the text words corresponding to each speaker.

30. The method of claim 29, further comprising distinguishing the text words corresponding to each speaker by presenting the text words corresponding to each speaker in different colors.